Lilly Tech Systems
Build an AI Email Assistant
Build a complete, AI-powered email assistant from scratch. You will connect to Gmail via OAuth, classify incoming emails by priority and category, generate context-aware draft replies with LLM, extract action items and summaries, and manage everything through a polished web dashboard — all in 6 hands-on steps.
What You Will Build
A fully functional AI email assistant that reads your Gmail inbox, classifies emails by priority and category, generates smart draft replies, extracts action items, and lets you review and send everything from a clean web dashboard.
Gmail Integration
Full OAuth 2.0 authentication with Gmail API. Fetch emails, parse threads, handle attachments, and manage labels programmatically.
Smart Classification
LLM-powered priority detection (urgent, high, normal, low), category tagging (meeting, task, FYI, personal), and sentiment analysis on every incoming email.
Draft Generation
Context-aware reply drafts that match your writing tone. Template system for common responses. Review, edit, and send directly from the dashboard.
Web Dashboard
A responsive email dashboard with priority inbox, draft review panel, action item tracker, and follow-up reminders — all built with Flask and vanilla JavaScript.
Tech Stack
Every component is open source or has a generous free tier. Total cost to run: $0 for development, under $3/month in production.
Python 3.11+
The core language for the backend, Gmail integration, LLM calls, and the web server.
Gmail API
Google's official API for reading, labeling, and sending emails. OAuth 2.0 for secure authentication.
OpenAI API
gpt-4o-mini for classification, summarization, and draft generation. Under $0.15 per 1M tokens.
Flask
Lightweight web framework for the dashboard API and serving the frontend interface.
SQLite
Zero-configuration database for storing email metadata, classifications, drafts, and action items locally.
APScheduler
Background job scheduler for periodic email polling, follow-up reminders, and digest generation.
Prerequisites
Make sure you have these ready before starting.
Required
- Python 3.11 or higher
- A Google account with Gmail
- A Google Cloud project (free tier) for Gmail API credentials
- An OpenAI API key (get one at
platform.openai.com) - Basic Python knowledge (functions, classes, async/await)
- A terminal (bash, zsh, PowerShell, or CMD)
Helpful but Not Required
- Experience with Flask or FastAPI
- Familiarity with OAuth 2.0 flows
- Basic understanding of LLM prompt engineering
- HTML/CSS/JavaScript basics for the frontend
Build Steps
Follow these lessons in order. Each step builds on the previous one. By the end, you will have a fully functional AI email assistant.
1. Project Setup
Create the project structure, install dependencies, set up Gmail API credentials, configure OpenAI, and establish the database schema. You will have a working foundation by the end.
2. Email Integration
Connect to Gmail via OAuth 2.0, fetch emails with pagination, parse threads and headers, and store email metadata in SQLite. Full working Python code included.
3. Email Classification
Build an LLM-powered classification pipeline that detects priority levels, assigns category tags, and analyzes sentiment for every incoming email.
4. Draft Generation
Generate context-aware reply drafts that match your writing tone. Build a template system for common responses and implement tone-matching with few-shot examples.
5. Smart Features
Add email summarization, action item extraction, follow-up reminders, and daily digest generation. Turn your inbox into an organized task list.
6. Web Interface
Build a responsive email dashboard with a priority inbox view, draft review/edit panel, action item tracker, and one-click send workflow.
7. Enhancements
Add multi-account support, Outlook integration, email scheduling, privacy controls, and production deployment. Includes a comprehensive FAQ.